Large language models pass a standard three-party Turing test
Proceedings of the National Academy of Sciences · 26.05.2026 · 123 (21)
DOI: 10.1073/pnas.2524472123 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 19.05.2026. Verfasst von Cameron R Jones und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Cognitive Science, University of California San Diego, La Jolla, CA 92093. Hauptthemen laut MEDLINE: Artificial Intelligence. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 67 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Open Philanthropy Project.
Abstract (Originalsprache)
The Turing test has been widely discussed as a test of machine intelligence, but it also provides a measure of how humans distinguish other humans from machines. We evaluated 4 systems (ELIZA, GPT-4o, LLaMa-3.1-405B, and GPT-4.5) in two randomized, controlled, and preregistered Turing tests on independent populations. Participants had 5 min conversations simultaneously with another human participant and one of these systems before judging which conversational partner they thought was human. When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. LLaMa-3.1, with the same prompt, was judged to be the human 56% of the time—not significantly more or less often than the humans it was being compared to. Without these prompts, however, the same models performed significantly worse (38% and 36%), and did not consistently outperform baseline models, ELIZA and GPT-4o (23% and 21%, respectively). A third study replicated these results in 15-min games: two PERSONA-prompted models achieved pass rates of 56% and 59%. The results constitute empirical evidence that artificial systems can pass a standard three-party Turing test. Interrogators’ reasoning focused more on stylistic and socio-emotional aspects of human behavior rather than more traditional notions of intelligence. The results have implications for debates about what kind of intelligence is exhibited by large language models, the social impacts these systems are likely to have, and the aspects of human behavior that people continue to see as unique.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Cognitive Science, University of California San Diego, La Jolla, CA 92093.
- Department of Psychology, Stony Brook University, Stony Brook, NY 11794.
ORCID-Profile: Cameron R Jones, Benjamin K Bergen
Förderung
- Open Philanthropy Project
Zitieren
APA
Jones, C. R., & Bergen, B. K. (2026). Large language models pass a standard three-party Turing test. Proceedings of the National Academy of Sciences, 123(21). https://doi.org/10.1073/pnas.2524472123
Vancouver
Jones CR, Bergen BK. Large language models pass a standard three-party Turing test. Proceedings of the National Academy of Sciences. 2026;123(21). doi:10.1073/pnas.2524472123
BibTeX
@article{Jones2026,
author = {Jones, Cameron R and Bergen, Benjamin K},
title = {Large language models pass a standard three-party Turing test},
journal = {Proceedings of the National Academy of Sciences},
year = {2026},
volume = {123},
number = {21},
doi = {10.1073/pnas.2524472123}
}
EUVIMED